Upcoming ELNA SNS Decentralization Sale

ELNA.ai is planning to launch its SNS DAO with a proposal to start the decentralization sale soon.

This post is a forum for the IC community to discuss the ELNA.ai SNS and ask the founding dev team any questions you may have.

ELNA’s mission is to democratize AI. Being the world’s first fully on-chain decentralized generative AI platform, ELNA allows for training of large language models to create AI chat agents that seamlessly integrate on any platform. Being built on the Internet Computer, ELNA’s vector db implementation in canisters allow for scalability and speed of data/code read and retrieval. And its RAG (Retrieval Augmented Generation) delivers highly contextualized and recent information that can be fact-checked using source citations.

Please refer to the Tokenomics and ELNA’s Whitepaper for further details.

Application: http://dapp.elna.ai/
Roadmap : https://www.elna.ai/roadmaps.html
Tokenomics : Sheet
More Info
Whitepaper : https://docs.elna.ai
Github: https://github.com/elna-ai
Dfinity Forum Posts : https://forum.dfinity.org/u/branbuilder/
Twitter : https://twitter.com/elna_live

It is recommended to review these documents before asking questions, as they may already contain answers.

10 Likes

Hey, at this point I think it might be premature for the ELNA.ai project to start an SNS sale. The project is not yet fully developed and I believe you should develop its fundamentals further before rushing it into a sale. This approach ensures a more robust and successful business in the long run.

1 Like

Hi @branbuilder

it seems that you are using the legacy SNS tokenomics excel sheet.

Please have a look at this SNS tokenomics wiki page, to get pointers to the most recent SNS init file template and also the SNS tokenomics analyzer.

3 Likes

Hi, this is a different topic so if more assistance is needed please create a new one to get more support.

Otherwise, in general you need to click the Start Dissolving in the Locked Neuron and then wait for the Dissolve Delay period - after it passes you will be able to Disburse your ICP:
image

Another option which could be faster (but you might need to give some discount) is to sell your II (includes the locked Neuron(s)) on https://idgeek.app

Hope this helps.

I thought you just received a developer grant.

Did you already complete that?

2 Likes

btw…
Don’t get me wrong. I am bullish on the ELNA project and vision, and hope it succeeds.

I am just very interested to try out your RAG db running on the IC.

1 Like

There is no consumer product yet so yeah I agree with you. I think the majority of projects who launched so far have not been ready and it showed with the lack of liquidity as they need the funds to develop. SNS should be used for complete or almost complete products.

3 Likes

Do you have a timeline for the roadmap items? Do you have a timeline for when you want to make an SNS proposal?

1 Like

Heyy @jennifertran! Do check our roadmap here:
https://www.elna.ai/roadmaps.html

And we’ll be announcing the exact SNS date soon.

2 Likes

Hi @Phasma, we appreciate you, but we believe there’s been a gap in understanding here.
ELNA’s functionality is already in full bloom, and our current work is geared mainly towards adding more agents to the current list -

Agents: :link:https://gpdbs-xqaaa-aaaah-adtiq-cai.icp0.io/

as well as releasing specialized agents as products that solve for societal problem statements - the first one in the works being an Ederly Care Companion to help alleviate their loneliness.

We hope you check the link out and ping us on Discord to be whitelisted and start training ELNA with your data to create agents of your liking and monetize it in the near future🫶

Discord: :link:ELNA

2 Likes

Hi @Sormarler, we appreciate you, but we believe there’s been a gap in understanding here.
ELNA’s functionality is already in full bloom, and our current work is geared mainly towards adding more agents to the current list -

Agents: :link:https://gpdbs-xqaaa-aaaah-adtiq-cai.icp0.io/

as well as releasing specialized agents as products that solve for societal problem statements - the first one in the works being an Ederly Care Companion to help alleviate their loneliness.

We hope you check the link out and ping us on Discord to be whitelisted and start training ELNA with your data to create agents of your liking and monetize it in the near future🫶

Discord: :link:ELNA

Thank you! Do you have a timeline for when the feature items on the roadmap will be completed?

1 Like

Hi, please can you clarify why your agent (Pratheesh) reports that is developed by OpenAI and identifies itself as ChatGPT?

I didn’t see such information mentioned in the Whitepaper but I had just a quick walk through so could have missed it (sorry in such case).
Also note that I’m not AI expert so I’m not making any conclusions.


Another test, this time trying Philomina - AI language model created by OpenAI, not running on ICP:

2 Likes

@branbuilder ,

These experiments from @plsak demonstrate you’re running most of your services off-chain, which is to be expected, because the IC does not yet offer the infrastructure to run everything you’re providing on-chain. The vector DB might be the exception, and I am very interested to learn more about your approach.

It would be awesome if you could join the Technical Working Group DeAI meetings. We meet every Thursday, and today we had a great discussion on the most pressing needs to make on chain AI a possibility.

Your input and sharing of experiences would be very valuable for the community. :pray:

7 Likes

Thanks @bjoernek
Will check the same and will do the needful

HI @icpp

We have completed 2 milestones and now are working on the on-chain Vector DB.

Well-wishers like you are our main source of motivation

@icpp
Your wonderful implementation of “Llama2.c” has been one of our inspirations to make an LLM agent on ICP, so that people can use implementations like the Llama2.c to create LLM agents with the help of RAG and other integration tools running on the ICP chain.

We have completed the initial orchestration and prompt engines, now testing the embedding and vector. Would love to have a second call with you for the integration of Llama2.c to ELNA.

Now there are memory and instruction limits to run larger models but we are confident of running it in future updates.

Llama2.c + ELNA = Chat GPT agents killer.

1 Like

Hi @jennifertran

We’ve completed the orchestration, prompt engine and integration of external LLMs like Chat GPT.

And we are in the process of completing our vector db by next month.

We’ve been working with multiple clients for various feature add-ons like twitter integration, telegram, chrome extension support etc.

2 Likes

Hi @plsak ,
Apologies for the confusion

ELNA is an LLM agent creation platform, meaning -

We help you create custom autonomous agents, using LLMs like Chat GPT (with prompt engine, vector db for knowledge base, orchestration etc.), which will be running on chain.

We are not building LLMs like GPT but support the integration of LLMs like GPT, Llama2.c etc.

The current bot which you tried is the integration of the OpenAI api, currently there are some limits to run larger models completely on the IC chain.

Hope we’ve cleared your doubts, please feel free to ask more for any clarity

4 Likes

Will join Technical Working Group DeAI from the next one, we had briefly popped in for the first one.

WAGMI possible - decentralised AI on ICP

As I have answered @plsak - our vision is to make autonomous LLM agents on IC.

3 Likes